Knowledge Planet: Old Dock's Tech Chronicles Zhihu: SamDeepThinking A technical manager who still codes heavily on the front line. From junior developer to tech lead, then tech manager, now leading the whole front‑ and back‑end development team—leveling up along the way. I have some insights on programming, career development, and tech management.
The article examines Elon Musk’s terse internal memo that forces managers to either explain a mistake, seek clarification, or act, using it to illustrate how high‑performing organizations cultivate judgment and responsibility rather than relying on layered approvals that stall execution.
The article presents a classic interview problem of locating the unique element in an integer array where every other value appears twice, and walks through three Java 17 implementations—brute‑force double loop, HashMap counting, and a bitwise XOR trick—detailing their time and space trade‑offs and practical considerations.
This article explains how to implement breadth‑first (BFS) and depth‑first (DFS) traversals—including level‑order, preorder, inorder, and postorder—on a binary tree in Java 17, covering both recursive and iterative approaches, performance considerations, and a lazy Iterable wrapper.
When a company’s internal tools each required separate logins, the team built an enterprise‑grade CAS single sign‑on system that unified identity, login entry, logout handling, and extensible integration, and they document the entire process across nine detailed articles.
The article explains why placing column C immediately after the equality column A in a composite index (A, C, B) avoids filesort and keeps performance stable regardless of how many values appear in the B IN list, outperforming other index orders such as (A, B, C).
This article walks through creating a minimal Java 17 RPC framework—from defining a simple binary protocol and using JDK serialization to implementing server loops, dynamic proxy‑based clients, and a runnable demo—while also highlighting the gaps that production‑grade solutions like Dubbo or gRPC must fill.
The article examines why the common update‑database, delete‑cache, add‑TTL workflow can still produce permanent stale data under high concurrency, explains the underlying race conditions, and compares several alternative strategies—including delete‑first, binlog‑driven invalidation, and lease‑based approaches—to help engineers choose the most reliable and low‑complexity solution for cache consistency.
The article argues that technical growth comes from shedding absolute beliefs, prioritizing sound architecture, writing clear code, avoiding needless hype, and applying best‑practice principles only where they truly fit, all backed by concrete development experiences.
The author argues that a technical manager should devote about 80% of their time to coding and only 20% to management, warning that abandoning hands‑on development erodes technical judgment, team trust, and long‑term career competitiveness.
The article explains that interviewers ask about the most difficult problem you solved to evaluate your analytical thinking, decision‑making process, and ability to handle uncertainty, and it offers guidance on how to describe the problem, solution steps, and trade‑offs effectively.
